What a Job Actually Costs You

Busy and profitable are not the same thing

Almost every trade business tracks revenue, because it’s the number that arrives in the bank and the number the invoices already add up to. Far fewer track what any individual job actually cost to deliver — and that’s the number that decides whether a busy year is also a good one.

The uncomfortable version of this: it is entirely possible to be fully booked, turn over more than last year, work more Saturdays than last year, and end up with less money. That happens when the mix of work shifts toward jobs that look fine on the invoice and lose money in the doing. Without per-job costs, there is no way to see it happening — the monthly totals still go up.

The four costs that hit every job

Materials are the obvious one, and the easiest to capture, because there’s a receipt. The trap isn’t knowing what you spent; it’s knowing which job you spent it on. A week of receipts in the truck door is a week of costs you can’t assign to anything.

Labour means anyone you paid to be there, including subcontractors — and, critically, including yourself. More on that below, because it’s the one that gets skipped.

Vehicle and travel is the cost people most often treat as fixed when it’s actually per-job. Fuel, wear, and the hour of driving that a far job costs and a near one doesn’t are real differences between two jobs at the same price.

Overhead — insurance, phone, software, accountant, storage — is genuinely fixed, but it still has to be paid out of the margin your jobs produce. A rough monthly figure divided by how many jobs you typically do in a month is enough to sanity-check whether a job type actually clears it.

Pay yourself a real wage, at least on paper

The single most common job-costing mistake in a one-person trade business is costing your own hours at zero. It makes every job look profitable, because the largest input is free. It also makes it impossible to compare two jobs honestly: the one that took you eleven hours and the one that took you three look identical if your time isn’t in the numbers.

Pick a realistic hourly figure for your own labour — what you’d have to pay someone competent to do that work — and put it into the cost of every job, even though no money changes hands. You’re not trying to produce accounts; you’re trying to produce a comparison. A job that only "makes money" because you worked for nothing is a job that needs repricing, and you can’t see that until your time has a number attached.

Track costs per job, not per month

Monthly totals are how most trades look at costs, and they hide exactly the thing you need to see. If materials were $4,100 across nineteen jobs, you know the month was expensive. You don’t know that three of those jobs ate a third of it, or which three.

Per-job costing answers a different and much more useful question: not "did we do well last month" but "which kind of work should we do more of." The answer is often surprising. The big prestigious job can turn out to be the one carried by a heap of unbilled hours, while the small repeat work nobody brags about quietly funds the business.

The receipt problem is the whole problem

In practice, job costing almost never fails because the maths is hard. It fails because the inputs never get captured. The receipt goes in a pocket, the pocket goes in the wash, and three supply runs across two jobs become one vague memory of "a lot at the merchant this week."

Whatever system you use, the test is whether capturing a cost takes seconds at the counter rather than minutes at a desk later. If it takes minutes later, it will not happen consistently, and inconsistent cost data is arguably worse than none — it makes some jobs look cheaper than they were purely because their receipts went missing.

What to actually do with the number

Job costing is only worth the effort if it changes a decision. Three it should change:

Reprice the work that doesn’t clear. If a job type consistently returns a thin margin once your own time is counted, the rate is wrong — not the customer. Raising a rate on a category of work is far easier than finding new customers to subsidise it.

Fix the estimate, not just the price. If material overruns are the problem rather than the rate, the fix is in how you estimate quantities — waste factors, overage, the second trip you always end up making — not in charging more for a number you keep getting wrong.

Say no to a category. Sometimes the honest answer is that a type of job doesn’t work at any price a customer will pay. Knowing that with numbers behind it is much easier than knowing it as a vague feeling that those jobs are annoying.

Where this gets easier

Voiczer is built so the costing happens as a by-product of the job rather than as a second admin task: log a cost by saying it, photograph a receipt and have the amount, supplier and date read off it automatically, and see revenue, cost and margin per job without assembling anything. The point isn’t the reporting — it’s that the numbers are actually there to report on, because capturing them never required an evening at the kitchen table.